1. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
Crop Yield Forecasting by Adaptive Neuro Fuzzy
Inference System
Pankaj Kumar*
VCSG College of Horticulture, GBPUA&T, Pantnagar, Uttarakhand
*pankaj591@gmail.com
Abstract
Meteorological uncertainties affect crop yield portentously during different stages of crop growing seasons,
therefore several studies have been carried out to forecast crop yield using climatic parameters with
empirical statistical regression equations relating regional yield with predictor variables. In this study an
attempt has been made to develop Crop Yield Forecasting models to map relation between climatic data
and crop yield. Present study was undertaken for forecasting rice yield by adaptive neuro fuzzy inference
system (ANFIS) technique based on time series data of 27 years, yield and weather data (w.e.f. 1981-82 to
2007-08) obtained from G. B. Pant University of Agriculture and Technology, Pantnagar, District Udham
Singh Nagar, Uttarakhand, India.
Keywords: climate, crop yield, ANFIS, forecasting
1. Introduction
In India it is possible to cultivate large number of crops due to diverse climatic conditions. Among
these crops Rice (Oryza sativa L.) is important food crops of the country. The total area under cultivation
of Rice 43.6 (22.6% of gross cropped area) and the total production of Rice is 94.1 million tonnes during
the year 2007-08 constituting about 42.9% of total food production (Economic Survey 2007-2008). Crop
failure on account of drought or flood will have a severe repercussion not only on the country’s economy
but also on food security.
The forecasting of crop yield may be done by using three major objective methods (i) biometrical
characteristics (ii) weather variables and (iii) agricultural inputs (Agrawal et al. 2001). These approaches
can be used individually or in combination to give a composite model. Forecasting is an significant aid in
effective and efficient planning and is more important aspect for a developing economy such as ours so that
adequate planning exercise is undertaken for sustainable growth, overall development and poverty
alleviation. In management and administrative situations the need for planning is great because the lead
time for decision making ranges from several years (for capital investments) to a few days or hours (for
transportation or production schedules) to a few seconds (for telecommunication routing or electrical utility
loading). Forecast of crop yield is of immense utility to the government and planners in formulation and
implementation of various policies relating to food procurement, storage, distribution, price, import-export
etc.
Several studies have been carried out to forecast crop yield using weather parameters (Huda et al.
(1975), Choudhary and Sarkar (1981), Kokate et al. (2000), etc.). However, such forecast studies based on
statistical models need to be done on continuing basis and for different agro-climatic zones, due to visible
effects of changing environmental conditions and weather shifts at different locations and areas. Therefore,
there is a need to develop area specific forecasting models based on time series data to help the policy
makers for taking effective decisions to counter adverse situations in food production.
In recent years soft computing techniques like, artificial neural network (ANN), fuzzy logic,
genetic algorithm and chaos theory have been widely applied in the sphere of prediction and time series
modelling. Adaptive neuro-fuzzy inference system (ANFIS) which is integration of neural networks and
fuzzy logic has the potential to capture the benefits of both these fields in a single framework. ANFIS
utilizes linguistic information from the fuzzy logic as well learning capability of an ANN. Thus the present
1

2. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
study was carried out to develop forecasting model based on time series agricultural data related to yield of
rice crops based on weather parameters.
2. Materials and Methods
2.1 Adaptive Neuro-Fuzzy Inference Systems (ANFIS)
Adaptive Neuro Fuzzy Inference System (ANFIS) is a fuzzy mapping algorithm that is based on Tagaki-
Sugeno-Kang (TSK) fuzzy inference system (Jang et al., 1997 and Loukas, 2001).ANFIS is integration of
neural networks and fuzzy logic and have the potential to capture the benefits of both these fields in a
single framework. ANFIS utilizes linguistic information from the fuzzy logic as well learning capability of
an ANN for automatic fuzzy if-then rule generation and parameter optimization.
A conceptual ANFIS consists of five components: inputs and output database, a Fuzzy system
generator, a Fuzzy Inference System (FIS), and an Adaptive Neural Network. The Sugeno- type Fuzzy
Inference System, (Takagi and Sugeno, 1985) which is the combination of a FIS and an Adaptive Neural
Network, was used in this study for rainfall-runoff modeling. The optimization method used is hybrid
learning algorithms.
For a first-order Sugeno model, a common rule set with two fuzzy if-then rules is as follows:
Rule 1: If x1 is A1 and x2 is B1, then f1 = a1 x1+b1 x2 + c1.
Rule 2: If x1 is A2 and x2 is B2, then f2 = a2 x1+b2 x2 + c2.
where, x1 and x2 are the crisp inputs to the node and A1, B1, A2, B2 are fuzzy sets, ai, bi and ci (i = 1, 2) are
the coefficients of the first-order polynomial linear functions. Structure of a two-input first-order Sugeno
fuzzy model with two rules is shown in Figure 1 It is possible to assign a different weight to each rule
based on the structure of the system, where, weights w1 and w2 are assigned to rules 1 and 2 respectively.
and f = weighted average
The ANFIS consists of five layers (Jang, 1993), shown in Figure 1. The five layers of model are as follows:
Layer1: Each node output in this layer is fuzzified by membership grade of a fuzzy
set corresponding to each input.
O1,i = µAi (x1) i = 1, 2
or
O1,i = µBi-2 (x2) i = 3, 4
(1)
Where, x1 and x2 are the inputs to node i (i = 1, 2 for x1 and i = 3, 4 for x2) and x1 (or
x2) is the input to the ith node and Ai (or Bi-2) is a fuzzy label.
Layer 2: Each node output in this layer represents the firing strength of a rule, which performs fuzzy, AND
operation. Each node in this layer, labeled Π, is a stable node which multiplies incoming signals and sends
the product out.
O2,i = Wi = µAi (x1) µBi (x2) i = 1, 2
(2)
Layer 3: Each node output in this layer is the normalized value of layer 2, i.e., the normalized firing
strengths.
O3i = Wi = w i i =1, 2
w +w
1 2 (3)
Layer 4: Each node output in this layer is the normalized value of each fuzzy rule. The nodes in this layer
are adaptive. Here Wi is the output of layer 3, and {ai,bi,ci} are the parameter set. Parameters of this layer
are referred to as consequence or output parameters.
2

3. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
O4i = Wi f i = Wi (ai x1 + bi x2 + ci ) i = 1, 2 (4)
Layer 5: The node output in this layer is the overall output of the system, which is the summation of all
coming signals.
2
2
∑W i fi
(5)
Y = ∑W1 i fi = 1
2
∑W 1
i
Figure 1.ANFIS architecture
In this way the input vector was fed through the network layer by layer. The two major phases for
implementing the ANFIS for applications are the structure identification phase and the parameter
identification phase. The structure identification phase involves finding a suitable number of fuzzy rules
and fuzzy sets and a proper partition feature space. The parameter identification phase involves the
adjustment of the premise and consequence parameters of the system.
Optimizing the values of the adaptive parameters is of vital importance for the performance of
the adaptive system. Jang et al. (1997) developed a hybrid learning algorithm for ANFIS to approximate
the precise value of the model parameters. The hybrid algorithm, which is a combination of gradient
descent and the least-squares method, consists of two alternating phases: (1) in the backward pass, the error
signals recursively propagated backwards and the premise parameters are updated by gradient descent, and
(2) least squares method finds a proper set of consequent parameters (Jang et al., 1997). In premise
parameters set for a given fixed values, the overall output can be expressed as a linear combination of the
consequent parameters.
AX = B
(6)
Where, X is an unknown vector whose elements are the consequent parameters. A least squares estimator of
X, namely X*, is chosen to minimize the squared error ‖ ‖ . Sequential formulas are employed to
compute the least squares estimator of X. For given fixed values of premise parameters, the estimated
consequent parameters are known to be globally optimal.
2.2 Study Area and Model Application
3

4. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
2.2.1 Study area
The present study was carried out to develop forecasting models for predicting the yield of rice
and wheat at Govind Ballabh Pant University of Agriculture and Technology, Pantnagar, District Udham
Singh Nagar, Uttarakhand, India, which is situated in the Shivalik range of Himalayan foot hills at
approximately 243.89 meters above mean sea level. It is located at 290 N latitude and 79.30 E longitudes.
This region is known as Tarai (meaning wet) and has a dry season from early October to mid-June and a
wet season from June to early October. The soil of Tarai is mainly silt and loamy having good moisture
storage capacity, about 7.3 pH and is highly productive. The annual rainfall of this region is about 1400
mm, which is subjected to large variation.
Yearly yield data of rice (q/ha) for 27 years (w.e.f. 1981-82 to 2007-08) was collected from the
University Farm of G. B. Pant University of Agriculture and Technology, Pantnagar, which is a
commercial well organized farm having 2714.43 ha land utilized for seed production, research purpose and
commercial farming. The farm is well irrigated, fully mechanized and well equipped with all modern
agricultural technologies and is governed with a high level management practices with the regular use of
high yielding varieties, recommended doses of fertilizers and manures, proper plant protection measures
including soil water management.
The time series weekly weather data of 27 years (from 1981-82 to 2007-08) were collected from
Agro-meteorological observatory situated at the crop research centre of G. B. Pant University of
Agriculture and Technology, Pantnagar, which is equipped with modern technology and instruments. Six
weather parameters were included in this study; namely average weekly temperature (T°C), average weekly
relative humidity (Rh%), average weekly sunshine hrs (S), average weekly total rainfall (P ), average
weekly number of rainfall days (n) and average weekly pan evaporation (E). However, weekly weather data
related to kharif (the autumn harvest also known as the summer or monsoon crop in India) crop seasons
starting from a fortnight before sowing up to one month before harvest were utilized for the development of
models in the present study therefore, the weather data for rice crop (Kharif season), from May 21 (about a
fortnight before sowing) to October 22 (one month before harvest) in each year were employed. Data for
last one month of the crop season was excluded as to forecast crop yield at least one month before harvest.
2.3 Model Application
Total 27 year data are divided in two sets namely training (calibration) and testing (validation). Fig 2 and
Fig3 indicates graphical representation of rice crop yield and weather variables are shown in.21 years data
from 1981-82 to 2001-02 were used for calibration of model rest 6 years data from 2002-03 to 2007-08
were used for testing of model.
Among various inputs like average weekly temperature (T), average weekly relative humidity
(Rh%), average weekly sunshine hrs (S), average weekly total rainfall (P ), average weekly number of
rainfall days (n) and average weekly pan evaporation (E) weather parameters best parameter to forecast rice
crop yield were chosen by Gamma test. Gamma test is a technique for estimating the variance of the noise,
or the mean square error, that can be achieved without over fitting (Jones, 2004). It is used for evaluation of
the nonlinear correlation between random variables, namely, input and output pairs. Gamma test was
applied on all possible combinations of weather variables. Gamma test concluded that only the combination
of weekly pan evaporation (E) and average weekly temperature (T) are sufficient to predict rice crop yield.
Including average weekly relative humidity (Rh%), average weekly sunshine hrs (S), average weekly total
rainfall (P ) and average weekly number of rainfall days (n) cause complexity in model and yields a poorer
root mean square value between actual yield and predicted yield.
Combinations of weekly pan evaporation (E) and average weekly temperature (T) were considered as the
inputs to the model, and Yield of the current year was considered as the output. Input space partitioning for
model structure identification was done by grid partition. Hybrid learning algorithm was used to train the
models for runoff prediction. The optimal learning parameters were determined by trial and error (Kim et
al., 2002) for constant trapezoidal membership function. In order to choose better model among developed
models root mean square error was computed.
4

5. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
3. Result and Discussions
The study revealed that the least value of root mean square error were obtained for model with two inputs,
weather data of weekly pan evaporation and average weekly temperature for rice crop in Kharif season,
from May 21 (about a fortnight before sowing) to October 22 (one month before harvest) and output as rice
crop yield of the current year. Among triangular, bell shaped, trapezoidal and gaussian membership
function, the trapezoidal membership function was found most suitable for this study. 2 trapezoidal and
constant membership functions were used for 30 epochs. The observed and estimated crop yields are shown
in Fig 5 and Fig 6 for training and testing periods respectively. Values of root mean square error is 3.485 in
case of training period and 1.756 in case of testing.
Rice (q/ha)
50
45
40
Yield (q/ha)
35
30
25
20
15
10
5
0
Year
Figure 2. Rice yield (q/ha)
140
120
100
T (°C)
80 Rh (%)
60 P (mm)
S (Hours)
40
E (mm)
20 n (day)
0
Figure 3. Weather variable
5

6. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
Actual Yield Predicted yield
50
45
40
35
30
25
20
15
10
5
0
1981-82
1982-83
1983-84
1984-85
1985-86
1986-87
1987-88
1988-89
1989-90
1990-91
1991-92
1992-93
1993-94
1994-95
1995-96
1996-97
1997-98
1998-99
1999-00
2000-01
2001-02
Figure 4.Observed and estimated rice crop yield during training period
Actual Yield Predicted Yield
45
40
35
30
25
20
15
10
5
0
2002-03 2003-04 2004-05 2005-06 2006-07 2007-08
Figure 5. Observed and estimated rice crop yield during testing period
6

7. Mathematical Theory and Modeling www.iiste.org
ISSN 2224-5804 (Paper) ISSN 2225-0522 (Online)
Vol.1, No.3, 2011
4. Conclusions
The present study discusses the application and usefulness of adaptive neuro fuzzy inference system based
forecasting approach for forecasting of rice crop yield. The visual observation based on the graphical
comparison between observed and predicted values and the qualitative performance assessment of the
model indicates that ANFIS can be used effectively for crop yield forecasting.
References
Jang, J.-S. R. (1993) “ANFIS: adaptive network-based fuzzy inference systems”, IEEE Trans. Syst., Man
Cybern 23, 665–685.
Jang, J.S.R.; Sun, C.T. and Mizutani, E.(1997), “Neuro-Fuzzy and Soft Computing, A Computational
Approach to Learning and Machine Intelligence”, Prentice Hall, NJ,USA ISBN 0-13-261066-3.
Kim, B.; Park, J.H.; Kim, B.S.(2002), “Fuzzy logic model of Langmuir probe discharge data”, Comput
Chem,26(6), 573–581.
Loukas, Y.L (2001) “Adaptive neuro-fuzzy inference system: an instant and architecture-free predictor for
improved QSAR studies”,J Med Chem 44(17),2772–2783.
Nayak, P.C., Sudheer, K.P., Rangan, D.M. and Ramasastri, K.S. (2004) “A neuro-fuzzy computing
technique for modelling hydrological time series”, J. Hydrology 291, 52-66.
Takagi, T., M. Sugeno.(1985) “Fuzzy identification of systems and its application to modeling and
control”, IEEE Transactions on Systems, Man, and Cybernetics, 15, 116–132.
Zadeh, L.A. (1965)”Fuzzy Sets”, Information and Control 8, 338-353.
Agrawal; R.; Jain, R.C., Mehta, S.C. (2001) “Yield forecast based on weather variables and agricultural
inputs on agro-climatic zone basis”, Ind. J. Agri. Sci. 71 (7), 487-490.
Huda, A.K.S.; B.P. Ghildyal; Tomar, V.S., Jain, R.C. (1975) “Contribution of climate variables in
predicting rice yield”, Agric. Met. 15,71-86.
Kokate, K.D.; Sananse, S.L. , . Kadam, J.R. (2000) “Pre-harvest forecasting of rice yield in Konkan region
of Maharashtra” J. Maha. Agric. Univ. 25 (3),289-293.
Chowdhury, A., Sarkar, M.B. (1981) “Estimation of rice yield through weather factors in a dry sub-humid
region” Mausam 32 (4), 393-396.
Jones, A. J.(2004) “New Tools in Non-linear Modeling and Prediction”, Computational Management
Science1(2), 109-149.
7

8. International Journals Call for Paper
The IISTE, a U.S. publisher, is currently hosting the academic journals listed below. The peer review process of the following journals
usually takes LESS THAN 14 business days and IISTE usually publishes a qualified article within 30 days. Authors should
send their full paper to the following email address. More information can be found in the IISTE website : www.iiste.org
Business, Economics, Finance and Management PAPER SUBMISSION EMAIL
European Journal of Business and Management EJBM@iiste.org
Research Journal of Finance and Accounting RJFA@iiste.org
Journal of Economics and Sustainable Development JESD@iiste.org
Information and Knowledge Management IKM@iiste.org
Developing Country Studies DCS@iiste.org
Industrial Engineering Letters IEL@iiste.org
Physical Sciences, Mathematics and Chemistry PAPER SUBMISSION EMAIL
Journal of Natural Sciences Research JNSR@iiste.org
Chemistry and Materials Research CMR@iiste.org
Mathematical Theory and Modeling MTM@iiste.org
Advances in Physics Theories and Applications APTA@iiste.org
Chemical and Process Engineering Research CPER@iiste.org
Engineering, Technology and Systems PAPER SUBMISSION EMAIL
Computer Engineering and Intelligent Systems CEIS@iiste.org
Innovative Systems Design and Engineering ISDE@iiste.org
Journal of Energy Technologies and Policy JETP@iiste.org
Information and Knowledge Management IKM@iiste.org
Control Theory and Informatics CTI@iiste.org
Journal of Information Engineering and Applications JIEA@iiste.org
Industrial Engineering Letters IEL@iiste.org
Network and Complex Systems NCS@iiste.org
Environment, Civil, Materials Sciences PAPER SUBMISSION EMAIL
Journal of Environment and Earth Science JEES@iiste.org
Civil and Environmental Research CER@iiste.org
Journal of Natural Sciences Research JNSR@iiste.org
Civil and Environmental Research CER@iiste.org
Life Science, Food and Medical Sciences PAPER SUBMISSION EMAIL
Journal of Natural Sciences Research JNSR@iiste.org
Journal of Biology, Agriculture and Healthcare JBAH@iiste.org
Food Science and Quality Management FSQM@iiste.org
Chemistry and Materials Research CMR@iiste.org
Education, and other Social Sciences PAPER SUBMISSION EMAIL
Journal of Education and Practice JEP@iiste.org
Journal of Law, Policy and Globalization JLPG@iiste.org Global knowledge sharing:
New Media and Mass Communication NMMC@iiste.org EBSCO, Index Copernicus, Ulrich's
Journal of Energy Technologies and Policy JETP@iiste.org Periodicals Directory, JournalTOCS, PKP
Historical Research Letter HRL@iiste.org Open Archives Harvester, Bielefeld
Academic Search Engine, Elektronische
Public Policy and Administration Research PPAR@iiste.org Zeitschriftenbibliothek EZB, Open J-Gate,
International Affairs and Global Strategy IAGS@iiste.org OCLC WorldCat, Universe Digtial Library ,
Research on Humanities and Social Sciences RHSS@iiste.org NewJour, Google Scholar.
Developing Country Studies DCS@iiste.org IISTE is member of CrossRef. All journals
Arts and Design Studies ADS@iiste.org have high IC Impact Factor Values (ICV).